1. Chapter three
3.0 METHODOLOGY
3.1 DATA ENVELOPING ANALYSIS
Data Enveloping Analysis is a non-parametric deterministic methodology that computes relative
efficiencies of entities, called decision making units (DMU), based on the empirical data that consist of
the combination of inputs and resulting outputs obtained for each DMU at a given technology.
3.2 DATA AND MODEL
Data used in the analysis consist of variables from 2001 to 2013 (inputs and output telecom variables)
data covering annual observations were primarily taken from ITU telecommunication database- 2013.
To measure efficiency in telecommunication industry, a DEA model was specified comprising of two
input variables: Total Investment in telecommunications and total number of staff (contract and full
time staff) employed by the telecommunication industry. Similarly studies by Koski, and Majumidar
(2011), Investment in telecommunication is used as a proxy for capital that is used to acquire, maintain
and upgrade the telecommunications infrastructure that supports the services rendered by the industry.
Choice of output variables was based on the interaction between firms and policy makers while for
firms, revenue is important to remain profitable and able to invest in advanced technologies; policy
makers are more geared towards goals such as affordability of the services and universal service, hence
telephone subscriptions and internet using (Giokas et al, 2008) are used for output choices.
Henceforth, the model consists of three output variables, namely total telephone subscriptions, internet
sers and revenue. Total telephone subscriptions include both fixed lines and mobile subscriptions.
Though fixed lines and mobile subscription could stand as two individual variables, this study followed
Lam and Shiu (2008) and Giokas et al (2008)’s procedure to reduce number of variables and allow
variability whenever study include a small number of DMUs.
While total telephone subscriptions include large portions of basic communication in Nigeria, the
number of internet users was used to reveal the level of adoption of advanced telecommunication
technologies and total revenue in the telecommunication industry (measured in millions USD) was used
as an indicator for the sector productivity which includes earnings from direct provisions of the
telecommunication services to public excluding revenue of resellers (ITU, 2003).
3.3 THE MATHEMATICAL FORMULATIONS OF DEA
The Linear programming Technique is used to find the set of coefficients (U’s and V’s) that will give the
highest possible efficiency ratio of outputs to inputs for the service units being evaluated.
In the model bellow

2. Maximize Ө =U1Y10 +U2Y20+……….+UrYro = Σ
V1X10
(i.e maximize the efficiency rating for service unit 0)
This is subject to the constraint that when the same set of U and V co efficient is applied to all other
service units being compared, no service unit (SU) will be more than 100% efficient as follows:
When U,…., U s > 0 and V1,…..> 0
SUj = Service unit number j
J = Number of service units (SU) being compared in the DEA analysis